[IBMiOSS] Weird `rowcount` value in ibm_db_dbi (Python 3)

In Python 3.6 for PASE, I'm using ibm_db_dbi and have a SELECT of the form

c1.execute('''
select count(*) from myfile where key = ?
''', (value,))
print(c1.fetchone()[0])

The keys are unique, and the above prints 1.

But when I try to update that row with

c1.execute('''
update myfile set field = ? where key = ? with nc
''', (value1, value2))
print(c1.rowcount)

it prints 8589934591. The whole file only has a few thousand records.

What the heck?

Kevin, if you're reading this, should I create an issue on GitHub? If
so, should it be on your IBM i fork?

John Y.